The intron/exon structure of ancient genes will be analysed in order to test the hypothesis that the original genes were pieced together from exons and thus that protein structures are built up in three dimensions from 'modules' that correspond to the exon products, which may then fold quasi-independently. Ancient genes are those for basic biochemical processes that evolved fully before the split into prokaryotes and eukaryotes. The genes for the glycolytic enzymes triosephosphate isomerase (TIM) and pyruvate kinase (PK) will be cloned from a wide variety of organisms: archaebacteria, fungi, and ancient and modern plants and animals in order to deduce a clear evolutionary tree of their intron/exon patterns. The goal is to develop a clear understanding of the course of evolution of gene structure over the development of the eukaryotes, the last billion and a half years, and to infer how the genes were assembled at the very beginning of evolution, three billion years ago. The demonstration that genes were put together from 'mini-genes' and hence that proteins were first put together from 'mini-proteins' will both change one's conception about the course and ease of evolution and restate the protein folding problem in a form easier of solution. To support this work, techniques for the still more rapid sequencing of DNA will be developed. The initial goal will be to increase the rate of sequence determination to a level of 10kb/week for a single worker. The next target will be to achieve a level of 50kb/week. This involves increasing the rate of sequencing by one to two orders of magnitude over the current technology. This will be done by applying the genomic sequencing techniques to the analysis of eukaryotic sequences cloned in cosmids and to bacterial sequences directly. The long term goal is to make it possible to sequence a cosmid in a week, and thus to bring the rate of sequence aquisition to the level of two megabases/person-year, so that single individual can sequence an entire bacterium. Such an increase in rate would profoundly change our knowledge of gene sequence and make it possible to sequence appreciable portions of the human genome.